Inside Anthropic's $2 Trillion Pitch

What we see in Anthropic's S-1 filing... A piece, not a say... The heart of the matter... Meta needs $95 billion... The 'AI super chip' briefing is Wednesday...

Anthropic's books are finally open...

This morning, AI startup Anthropic – the maker of the Claude models – filed its S-1 to go public on the Nasdaq under the ticker "ANT."

Goldman Sachs and Morgan Stanley lead a list of 22 underwriters.

An S-1 discloses a company's business operations, financials, risks, and leadership to investors. The U.S. Securities and Exchange Commission ("SEC") requires one from any company looking to go public or, in some cases, to sell additional shares.

We've been waiting on this S-1 for months. Anthropic filed a draft registration with the SEC back in June but kept pushing back its IPO plans. As we covered last Thursday, the Wall Street Journal got an early look at a draft.

In that issue, we quoted Stansberry Research senior analyst Josh Baylin about how developers are flocking to Anthropic's tools. And Josh named three things to look for in an upcoming S-1: audited revenue, the share count, and "lockup" terms.

Today delivered only the first (audited results through 2025). The share count was blank. And the precise lockup details – when early investors are able to sell their shares in public markets – remain unclear. But the filing says the company is planning to let early investors and Anthropic employees sell at least some of their shares as soon as the first day of trading.

Word is, though, that bankers are aiming to raise as much as $100 billion at a valuation near $2 trillion, which would top June's SpaceX (SPCX) debut as the biggest IPO ever.

Anthropic also hasn't set a date yet. But under SEC rules, its "roadshow" can start no sooner than 15 days after today's filing, so the IPO will happen no sooner than mid-November.

What we see...

This S-1 runs more than 300 pages, far too much to cover in one Digest. But here are the most important takeaways...

Anthropic's growth is real. In the first half of 2026, Anthropic reported $16.3 billion in revenue. That's up 1,300% from the same period last year and more than 3.5 times the company's revenue for all of 2025.

Anthropic also reportedly told prospective investors that third-quarter revenue hit $16.8 billion, which is roughly a $67 billion annual pace. About 80% of first-half sales came from customers paying by usage.

More than 1,900 customers now spend $1 million or more a year to use Anthropic's tools, including about 250 spending $10 million-plus.

The filing is less clear about how many individual subscribers it has overall and what kind of revenue it might generate from them.

The company is (again) warning about its own product. Anthropic CEO Dario Amodei has been one of the more vocal AI executives about the threat AI poses to the labor market – and, more widely, to humanity.

The company put some of that in writing here. From the S-1's overview of the AI landscape...

While the potential benefits of AI are extraordinary, its rapid advancement will impact the economy and society in ways that could create significant disruption, particularly if the technology is not developed and deployed responsibly. For example, bad actors have already conducted high-profile cyberattacks using AI’s capabilities to discover vulnerabilities, generate malware, and evade detection tools; other risks associated with misuse, including in the domain of biology, could result in even greater harm.

The S-1 form includes a "Risk Factors" section, which lays out every material risk a buyer should weigh. In Anthropic's case, that includes models behaving in "unintended, misaligned, or autonomous ways" and cases where "use of our AI models results in harm."

They're basically begging the government – "please, regulate us!" – and/or covering their backs: "Don't say we didn't warn you if AI agents start taking over the world."

Its $80.7 billion first-half loss isn't as bad as it looks, but it's still not great. On paper, Anthropic lost $80.7 billion in the first half of 2026. But $72.4 billion of that came from marking up convertible notes and similar instruments as the company's private valuation soared.

Convertible notes are loans that investors – like Amazon (AMZN), in Anthropic's case – can later swap for stock. Because lenders have a chance at that upside, the company pays little or no interest.

Here's the catch: The notes are treated as a debt that grows with the stock's value, so every jump in Anthropic's valuation shows up as a paper loss. The more the company is worth, the bigger the "loss" it has to book. But it's still a chunk of value that it owes eventually.

Meanwhile, operating losses weren't nearly as high. They were about $700 million in the first half of this year, which is much smaller than the $2.1 billion loss in the first half of last year.

Anthropic did report positive operating cash flow of $3.9 billion in the first half of this year, versus an operating cash flow deficit of about $1.5 billion in the first six months of 2025.

Altogether, Anthropic had more than $131 billion in cash and short-term investments at the end of June. It raised more than $90 billion in two private funding rounds this year, so that's not a surprise. Now it's turning to the public market.

You'd own a piece, not a say. Anthropic is a public benefit corporation ("PBC").

A five-member trust with no financial stake elects a majority of the board. Anthropic's seven co-founders get a single share carrying 50.1% of all shareholder votes. Public buyers get one vote per share, with no way to outvote either group.

And the board is already planning to dilute shareholders. As we wrote last week, Anthropic has decided to put 5% of its shares into a "public benefit fund," and the board can issue up to 2% more each time the stock climbs 50% from its last milestone, starting at the IPO price.

So, the better the stock does, the more shares get printed, and each share represents a smaller piece of the company.

The heart of the matter...

Financing.

Anthropic has years of bills ahead – and it needs to pay them.

Computing power (or "compute") and infrastructure cost Anthropic $8.9 billion in the first half, more than in all of 2025. And the S-1 says that as of June 30, the company owed about $265.6 billion in fixed or minimum compute payments, on contracts running as long as 20 years.

That doesn't even include its equipment leases – like one that costs it $125 billion over five years. Nor does it cover deals with no minimum spending amount, like one involving SpaceX data centers and Nvidia (NVDA) chips that could run as high as $120.5 billion.

Anthropic doesn't give a single total. But global news service Reuters, which saw an earlier draft of the S-1, says the company has at least $518 billion of commitments over roughly the next decade.

Meanwhile, Anthropic admits most of its customers can cut back any month, and roughly a fifth of revenue came from just two of them.

The S-1 also acknowledges risks to the AI boom that we've been writing about in these pages, like a shortage of available energy supply and growing public pushback against data centers.

So even with positive cash flow so far this year – and a potential $100 billion IPO haul – the company warns it may need to "seek additional financing sooner than we currently anticipate." It lists compute and infrastructure first among uses for the IPO money.

Anthropic isn't the only tech company scrambling for cash to feed its AI build-out...

As we wrote Thursday, Meta Platforms (META) said in its latest quarterly earnings filing that it could need to raise "hundreds of billions of dollars" over several years, starting as early as this fall.

And this morning, Meta announced plans for $35 billion in stock offerings. That's part of a plan to raise as much as $95 billion in equity, including a separate $60 billion program to sell its new nonvoting Class C shares over time.

Meta already has about $825 billion in future spending commitments, mostly for data centers and other AI infrastructure. And three of its own suppliers – Nvidia, Advanced Micro Devices (AMD), and Broadcom (AVGO) – said they're interested in buying up to $5 billion of the new shares.

What a tangled web they all weave... SpaceX, which supplies Anthropic with computing power, also raised its own public money in the largest IPO ever back in June. Its shares jumped 20% on Day 1... and later fell below the offering price, erasing more than $1 trillion in market value.

But back to the Anthropic IPO (and the next one like it from ChatGPT maker OpenAI)... At a $2 trillion valuation, buyers would pay about 30 times Anthropic's annualized revenue (as of the latest quarter) for a company with fixed bills in the hundreds of billions.

As Jeff Brown, the founder of our corporate affiliate Brownstone Research, wrote in the Masters Series just yesterday...

These two IPOs will be nothing more than liquidity events for the effective altruists, venture-capital firms, and private equity that invested in Anthropic and OpenAI as private companies.

They will make hundreds of billions of dollars and sell their shares happily to unsuspecting retail investors who don't understand the nuances of the strings that are being pulled in the background.

Investors will regret buying into these IPOs in the months that follow. As margins collapse in these two companies and the competitive market flourishes, they simply won't maintain the kind of market share that they have today.

So be warned – and be wise...

Our advice: Don't place an order for Anthropic shares on Day 1.

If you want to own it, at least wait for its first quarterly report as a public company and watch whether revenue keeps outrunning compute spending.

It's the same idea we shared Thursday: In the AI trade right now, align yourself with the cash. By buying into an IPO, you are the cash.

Now, that doesn't mean there aren't incredible investment opportunities related to these IPOs now that are worth pursuing. As we said, Anthropic is growing and has every intention of continuing to do so. So are its competitors.

Both Anthropic and OpenAI "will likely raise on the order of $200 billion between the two IPOs," Jeff wrote... They will "absolutely" hand that capital to other tech companies to fund the AI infrastructure build-out.

And this is where you want to look for opportunities today... Look for the companies that are less understood and garner much less attention in the financial media – yet are critical to the development of AI.

This is where smart money floods before a company becomes a household name – not into a mega-cap IPO selling event for early investors.

But smart investing doesn't have to mean cautious, low-upside plays...

Whitney Tilson and Jeff converge on the 'AI super chip'...

Later this week, Jeff – a Silicon Valley tech veteran – is sitting down on camera with our colleague Whitney Tilson to reveal his latest discovery. It's what they say could be the biggest investment opportunity of 2027.

Jeff recommended Nvidia before it soared 35,000%... and told folks to buy bitcoin before it rose 30,000%.

His subject now is a new AI "super chip" they say is 50 times faster than Nvidia's – slated for release soon, with the potential to ignite a new AI bull run.

Jeff says a tiny California company – plus five little-known stocks he calls the "Fabulous Five" – could hand investors 10 times their money.

Now, regular readers know that Whitney has expressed caution over AI becoming a "bubble." But he and Jeff agree there's still money to be made and ways to do it, even if the broad market struggles.

As Whitney wrote in his free daily letter today...

There will be a lot of pain in the AI sector when the bubble bursts.

But the strongest companies will not only survive but thrive... And I don't just mean the cash-gushing tech giants like Meta Platforms (META) and Microsoft (MSFT).

Join Whitney and Jeff this Wednesday to get the full story in a free presentation. They'll also share two free recommendations. The last time Whitney and Jeff gave away a free pick together, it quadrupled in a little more than a year.

Click here to reserve your free spot now.

Before we go, one non-AI note...

I opened up my account on Musk's X social platform this morning and was welcomed with this fear-provoking, surely AI-curated trio of news items about the war in Iran and Middle East tensions...

I'd say this list just about covers the wide range of possibilities about what might happen next in the Middle East. Like we've been saying, it looks like a major move in the war is coming soon.

AI aside, one way or another, it's wise to expect war-driven volatility ahead.

New 52-week highs (as of 10/23/26): Alpha Architect 1-3 Month Box Fund (BOXX), BP (BP), Maplebear (CART), Chord Energy (CHRD), Cintas (CTAS), Chevron (CVX), Dorchester Minerals (DMLP), EOG Resources (EOG), Equinor (EQNR), GitLab (GTLB), IQVIA (IQV), Kinder Morgan (KMI), Kimbell Royalty Partners (KRP), Cheniere Energy (LNG), Plains All American Pipeline (PAA), Qualys (QLYS), Roche (RHHBY), USCF SummerHaven Dynamic Commodity Strategy No K-1 Fund (SDCI), Saturn Oil & Gas (SOIL.TO), Veeva Systems (VEEV), State Street Energy Select Sector SPDR Fund (XLE), and ExxonMobil (XOM).

Do you have a comment or question for us? As always, e-mail us at feedback@stansberryresearch.com.

All the best,

Corey McLaughlin with Nick Koziol
Baltimore, Maryland
October 26, 2026